On improving laser data for extrinsic LRF/camera calibration

This paper proposes improvement of calibration data to be used in the extrinsic LRF/camera calibration problem. The proposed method reduces the mean of error between the acquired laser data and the real world feature that should be detected by the LRF. The key idea stems from the fact that an LRF has angular resolution that can be statistically improved. The method requires only simple operations. Comparisons in real world and in simulation are presented showing that projection error is significantly reduced.